Week 36 Day 6: The Wisdom and Madness of Crowds

In a jar-guessing contest (how many jelly beans?), the average guess of 100 independent people is remarkably accurate -- often better than any individual guess. This is crowd wisdom. But if everyone can see each other's guesses and copies the first few answers, the average becomes terrible. That is crowd madness. Markets work the same way.

When crowds are wise (conditions for market efficiency): (1) Independence: investors make decisions based on their own analysis, not by copying each other. (2) Diversity: different investors use different methods, time horizons, and information sources. Value investors, momentum traders, fundamental analysts, and quantitative algorithms all contribute diverse signals. (3) Decentralization: no single actor can dominate the market. (4) Aggregation: a mechanism (the market) efficiently combines all these diverse, independent signals into a price. When crowds go mad (conditions for bubbles and crashes): (1) Correlation: investors copy each other's trades (herding, algorithm mimicry, ETF flows that buy/sell all stocks simultaneously). (2) Homogeneity: everyone uses the same strategy, the same data, the same time horizon. (3) Contagion: panic or euphoria spreads faster than analysis. Social media accelerates contagion from days (1929 crash) to hours (2021 meme stocks) to minutes (flash crashes). (4) Leverage: borrowed money amplifies both the bubble and the crash, as margin calls force selling that triggers more margin calls. The implication for individual investors: over long time horizons (10+ years), the crowd is wise -- markets price assets reasonably and generate positive real returns. Over short time horizons (days to months), the crowd can be mad -- prices deviate from fundamentals due to herding, leverage, and emotional contagion. Strategy: invest for the long run (capture crowd wisdom), ignore the short run (avoid crowd madness).

Surowiecki (2004) formalized the conditions for crowd wisdom (diversity, independence, decentralization, aggregation) building on Galton's (1907) observation that the median estimate of an ox's weight at a county fair was within 1% of the actual weight. In financial markets, Hayek (1945) argued that prices are an information aggregation mechanism: the market price of a stock reflects the dispersed knowledge of millions of participants, each contributing a piece of the information mosaic. This is the theoretical foundation of the Efficient Market Hypothesis (Fama, 1970). However, Lo (2004) proposed the Adaptive Markets Hypothesis, which recognizes that the conditions for crowd wisdom (independence, diversity) vary over time: during calm markets, diverse investment strategies and independent decision-making produce efficient prices; during stress, herding, correlation, and homogeneity increase, reducing the crowd's 'effective diversity' and producing inefficient (and often extreme) prices. Danielsson, Shin, and Zigrand (2012) modeled the feedback loop between risk management and market dynamics: during stress, coordinated risk management (VaR-based selling, margin calls, stop losses) creates artificial correlation -- diverse portfolios that were independent during calm markets become identical during stress (everyone sells the same things simultaneously). This endogenous risk (risk created by the participants' response to risk) explains why market crashes are more severe than market rallies: the crash itself activates mechanisms (forced selling, margin calls) that amplify the crash. For passive investors, this framework justifies a long-term horizon: the wisdom-of-crowds component of market pricing (efficient in the long run) dominates the madness-of-crowds component (inefficient in the short run) over sufficiently long holding periods.